移动P2P分布式信任模型设计

移动P2P网络的开放性和松耦合特性使得节点恶意攻击行为普遍存在,而现有基于声誉的信任模型大都基于“信誉值高的节点评价推荐越可信”的假设,无法识别恶意节点动态策略性攻击行为。针对这一问题,将社会网络相关理论引入信任系统,提出一种基于社会距离的信任模型（SD^2Trust）。该模型区分了服务可信度和推荐评价可信度,用多维结构同型性描述向量刻画节点网络地位和行为特征,根据社会距离确定推荐节点集和推荐信誉计算权重,综合信任考虑了诋毁风险。理论分析和实验结果表明,该模型能有效对抗恶意节点动态策略攻击行为。     

一种基于名誉的P2P网络信任管理模型

由于P2P系统的开放、匿名等特点,节点间的信任关系往往很难通过客观的信任关系建立。本文构建了一种基于实体间交易历史和推荐的信任模型,提出了依据历史交易记录来计算直接信任值的方法,定义了信任推荐和信任聚合两种运算,运用分布式哈希表来处理信任值的存储和查询,并给出了信任咨询和更新的算法。实验表明,该模型可以提供较可靠的信任推荐服务,同时能较好地消除诋毁、恶意推荐等安全问题。     

Impact of dissipative heat and radiative heat on MHD viscous flow through a slandering stretching sheet with temperature-dependent variable viscosity

The purpose of the current investigation is to discuss the varying physical properties, that is, thermal conductivity and viscosity of a viscous time-independent flow past a slandering stretching surface. The heat transfer phenomenon is enhanced by the inclusion of an external heat source under the action of dissipative heat energy and thermal radiation. The implementation of convective boundary conditions affects the heat and solutal transfer phenomena with the inclusion of chemical reactions. The flow characteristic is directed by the system of differential equations that are converted to nonlinear ordinary type by appropriate assumptions of similarity variables and stream function. Furthermore, the “homotopy analysis method” is deployed to simplify the system. The convergence of the methodology is also discussed with their various order of approximations. The conformity of the solution is presented along with validation in tabular form. Moreover, the effect of the parameters on flow profiles is explained graphically, and the major contribution of few important parameters is laid down. The rate of shear stress along with the longitudinal velocity profile augments with the enhanced variable viscosity. Furthermore, the reverse impact is rendered for the augmented values of the variable viscosity.     

构造基于推荐的Peer-to-Peer环境下的Trust模型

在诸如文件共享等无中心的Peer-to-Peer环境中,资源共享是用户自愿的行为.在这类系统中,由于用户不为自身的行为担负(法律)责任,因而节点间的信任关系往往很难通过传统的信任机制建立.一种更合理的考虑是参考人际网络中基于推荐的信任关系建立方法.现有的模型不能很好地解决模型的迭代收敛性问题,同时缺乏对诸如冒名、诋毁等安全性问题的考虑.针对上述问题,在节点推荐的基础上提出了一种基于Peer-to-Peer环境的信任模型,并给出了该模型的数学分析和分布式实现方法.分析及仿真表明,该信任模型较已有模型在迭代的收敛性、模型的安全性等问题上有较大改进.